Johannes Nyström

Model-based robustness analysis is the strategy of assessing whether an ensemble of scientific models converges on an outcome, in order to support confidence in that outcome. While commonplace in many model-based sciences, the viability of the strategy is contested. In this article, I advance the debate in two steps. I first argue that Schupbach’s influential account, based on explanatory elimination, is limited in its ability to support strong confirmation. I then develop a novel account according to which robustness provides confirmation through assessments of the strength of underlying scientific constraints on model building. I also formally analyse the proposed epistemic mechanism in a Bayesian framework. The results show that it can generate both confirmation and disconfirmation under plausible conditions, and reveal what determines its significance. The account therefore provides a general framework for understanding the epistemic significance of robustness analysis in scientific modelling.

Colin Howson (2000) claims that the no miracles argument in favor of a realist interpretation of a scientific theory falls prey to the base rate fallacy and is therefore invalid on logical grounds. In response, Dawid and Hartmann (2018) claim that Howson only reconstructs a limited part of the argument. They argue that a more complete reconstruction of the no miracles argument takes into account the success frequency of a wider spectrum of scientific theory building, and therefore avoids the base rate fallacy. In a critical response to Dawid and Hartmann, Boge (2020) presents two challenges to their approach, both of which are designed to provide reasons for skepticism about treating observed success frequencies in science as connected to the relevant base rates. In this paper, I argue that Boge’s challenges are not effective.

The frequency of major theory change in natural science is rapidly decreasing. Sprenger and Hartmann (2019) claim that this observation can improve the justificatory basis of scientific realism, by way of what can be called a stability argument. By enriching the conceptual basis of Sprenger and Hartmann’s argument, this article shows that stability arguments pose a strong and novel challenge to scientific anti-realists. However, an anti-realist response to this challenge is also proposed. The resulting dialectic establishes a level of meaningful disagreement about the significance of stability arguments for scientific realism and indicates how the disagreement can ultimately be resolved.